NLR-TP-2000-618

Impact simulation of a frangibleImpact simulation of a frangibleImpact simulation of a frangibleImpact simulation of a frangibleapproach light structure by anapproach light structure by anapproach light structure by anapproach light structure by anaircraft wing sectionaircraft wing sectionaircraft wing sectionaircraft wing section

J.F.M. Wiggenraad, A. de Boer and R.H.W.M. Frijns

NationaalNationaalNationaalNationaal Lucht- en Ruimtevaartlaboratorium Lucht- en Ruimtevaartlaboratorium Lucht- en Ruimtevaartlaboratorium Lucht- en RuimtevaartlaboratoriumNational Aerospace Laboratory NLR

NLR-TP-2000-618

Impact simulation of a frangibleImpact simulation of a frangibleImpact simulation of a frangibleImpact simulation of a frangibleapproach light structure by anapproach light structure by anapproach light structure by anapproach light structure by anaircraft wing sectionaircraft wing sectionaircraft wing sectionaircraft wing section

J.F.M. Wiggenraad, A. de Boer and R.H.W.M. Frijns

This investigation has been carried out under a contract awarded by the NetherlandsDepartment of Civil Aviation, contract number 200.54.782.The Netherlands Department of Civil Aviation has granted NLR permission to publish thisreport.

This report is based on a presentation to be held at the third international KRASH users’seminar, January 8-10 2001, Arizona State University, Tempe, Arizona, USA.

The contents of this report may be cited on condition that full credit is given to NLR andthe authors.

Division: Structures and MaterialsIssued: 1 December 2000Classification of title: Unclassified

-2-NLR-TP-2000-618

Contents

Abstract 3

INTRODUCTION 3

EXEL TEST RESULTS 4

DRI/KRASH MODEL 5

ANALYSIS 6

CONCLUSIONS 7

ADDITIONAL RESULTS 8

ACKNOWLEDGEMENT 8

REFERENCES 8

12 Figures

(20 pages in total)

-3-NLR-TP-2000-618

Impact simulation of a frangibleapproach light structure by an

aircraft wing section

J.F.M. Wiggenraad, A. de Boer and R.H.W.M. FrijnsNational Aerospace Laboratory, The Netherlands

Abstract - A computer code has been developed to analyse the impact problem of a lightaircraft wing colliding with a frangible approach light structure. This code was evaluated onthe basis of a 3-dimensional model, representing such a structure, as produced by the Finishcompany EXEL. The numerical results were compared to full scale impact test results,provided by EXEL.

The failure mode and the impact peak force were well predicted. As the characteristics of thewing were not available in time (a wing section was unexpectedly made available byTransport Canada), the model was not yet accurate enough to predict the full force history.Hence, it was not possible to generate the impact energy for comparison with theexperimental value. This improvement will be carried out in the near future.

The merits of the new simulation capability were clearly shown. The capability provided bythe DRI/KRASH code is available on a pc, taking only a few minutes to run. In the nearfuture, the model of the EXEL mast will be improved as indicated above, and the validationof the code will be undertaken, by applying it to a different type of approach light system,made of aluminium.

INTRODUCTION

Certain types of equipment, such as approach light masts and ILS antennas, need to bepositioned close to runways at airports because of operational requirements. However, in caseof emergencies during take-off or landing, they become obstacles to flight safety. Hence, therequirements for the design of such structures are stringent and seem contradictory: they needto be stiff during operation, and collapse easily during an impact by an aircraft. Such acharacteristic is known as "frangible".

The design of frangible structures is not easy, because of the impact dynamics involved.Designers tend to consider the problem as quasi-static. The aviation authorities on the otherhand, have difficulty in evaluating the frangibility of structures proposed for installation atairports, and the International Civil Aviation Organisation (ICAO) is presently in the processof defining requirements and specifications. For the case of approach light structures, thecurrent approach to demonstrate the frangibility of a design is by means of a full-scale impacttest with a representative light aircraft wing section, at a speed of 140 km/h (Fig. 1).

In order to support the designer, as well as the certifying authorities or their consultants, acomputer code is being developed to predict the global behaviour of new approach lightdesigns when hit by the wing of a small aircraft at a relevant speed. This computer code,DRI/KRASH [1,2] was already available to analyse the crashing of aircraft structures andhelicopters into the ground. An aircraft model is built up of non-linear beams to represent the

-4-NLR-TP-2000-618

stiffness of the aircraft structure, mass points to represent the inertia of the structure, andenergy absorbing springs to represent the crushing part of the structure. The characteristics ofthe springs can be programmed according to test results, which are previously obtained forcomponents.

In order to allow the analysis of the impact of a wing with an approach light structure, it wasdecided to develop DRI/KRASH in such a way, that both the aircraft wing and the approachlight structure can be combined in a single model. The latest version of DRI/KRASH, asdeveloped specifically within the context of this project by DRI, contains a new option, whichallows the specification of different initial velocities for each of the individual masses. Thisfeature, along with many other improvements, extends the capability of the program to enablethe modelling of the coupled interaction between a moving mass system (aircraft) and astationary structural system (approach light structure). Both systems can now be modelledwith the beams, masses and springs available in KRASH.

The project to develop the computer code contains four phases: a feasibility phase, adevelopment phase, a validation phase and a phase to develop design criteria. The first twophases were completed with a midterm review and an update of the work programme. In aprevious paper, the results of the feasibility phase were presented [3]. A comparison wasmade of the numerical results obtained with a preliminary model with full-scale test results,generated with the commercially available frangible approach light structure of EXEL [4,5].EXEL is a Finish company, and has meanwhile supplied Amsterdam Airport with frangibleapproach light systems. Since then, structural tests were carried out on sections of the EXELmast [6], to generate failure criteria. Also, compression tests were carried out on the wingimpactors used in the tests, to establish their characteristics [7]. The wing impactors weremade available by Transport Canada, which participates in ICAO's Frangible Aids StudyGroup (FASG) with tests and computations.

The present report gives the final results of the development phase, in the form of animproved model to simulate the impact of EXEL's frangible approach light structure by anaircraft wing section. Validation of the computer code will be carried out by modelingapproach light structures of other manufacturers, for instance of the metal structures tested byTransport Canada.

EXEL TEST RESULTS

The impact test programme on six EXEL approach light masts implied the impact at140 km/h by a wing section (Fig. 2), representative of a light aircraft (as specified by theFASG), which was mounted on a truck. Forces, accelerations and the velocity of the truckwere measured, and high-speed video images were obtained.

The masts were approximately 6 m tall, and contained a dummy top mass of 15 kg,representative of three lights and the cross bar on which they are mounted. The masts areconfigured as a composite lattice structure (made of glass-fibre polyester) with a square crosssection of 400 mm x 400 mm, consisting of four vertical tubes, mounted on a steel bracket,and connected with diagonal tubes made of fiber-glass reinfored plastic. The masts wereimpacted at a height of 4.1 m above the ground.

-5-NLR-TP-2000-618

Impact forces were measured with two load cells, which were located between the wingbacking column and the cantilever beam of the support structure (and filtered with a 1 kHzlow pass filter), see figure 2. The total force was calculated as the sum of the two indications.The absolute maximum force during the impact was determined to be the peak force. Verticaland horizontal accelerations were measured also, to determine the relative motion between thesupport structure and the truck, but this information was less significant. The velocity of thetruck was measured by a pulse counter, which was mounted on the drive shaft of the truck.

The impact energy was determined by integrating the impact force with the forwarddisplacement of the impactor. This displacement was calculated from the recorded pulsecount versus time relationship. The correction, taking into account the movement of theimpactor relative to the truck body, was calculated by using the horizontal accelerationmeasured by the accelerometer.

The contact time was determined on the basis of the energy versus time curve. The rise timefrom the zero energy level to the maximum energy level was determined to be the contacttime.

Two natural frequencies were established: 184 Hz, the local resonance frequency of theimpactor, and 6.3 Hz, the global resonance frequency of the whole system.

The first three impact tests carried out by EXEL were representative of the conditions asmodelled with KRASH. The other three tests were more complicated, with lighting cablesconnected. The average results of the first three tests were: peak force 23.9 kN, impact energy36.7 kNm, and contact time 104 ms.

DRI/KRASH MODEL

The feasibility phase resulted in the development of a preliminary model [3]. The model was2-dimensional, based on general structural properties. The "frangibility" option (a movingwing section impacting a stationary approach light structure) was shown to act properly.Recommendations for improvement were made as follows:• Create a 3-dimensional model of the mast.• Incorporate structural failure criteria for the mast.• Refine crushing properties of the wing section.• Adjust the point of contact.• Update mass and stiffness properties of the mast.• Compare the overall deflection mode with test data.• Compare the force and energy levels with test data.• Compare the deformation and failure modes with test results.

In the current study, attention was paid to these recommendations. The preliminary modelwas extended to a 3-dimensional model as depicted in figures 3-4. In the preliminary model,the ‘moving’ wing was modelled with three beams and three masses, representing, the truckwhich moves the wing section, the wing backing column (steel beam) on which the wing wasmounted, and the wing itself, respectively. In the current model, the ‘moving’ wing isrepresented by two parallel series of three beams and three masses, each with half the stiffness

-6-NLR-TP-2000-618

and half the mass properties. In each series of three beams, the beam representing the wingsection has non-linear stiffness behaviour and has been tuned with results from the EXELtests [5]. At the time that the current study was carried out, the characteristics obtained fromthe static compression tests on the wing, performed at NLR [7], were not yet available.

The wing model is connected with both the two vertical front tubes and the two vertical afttubes. The stiffness of the beams, which represent the connection between the wing and thetwo aft tubes, becomes ‘active’ when the front tubes are deformed in such a way that they‘make contact’ with the aft tubes.

The top crossbar with three lights is represented by a lateral beam, which connects the twoframe halves in the XZ-plane with each other, as depicted in figure 4. The total top mass of15 kg, representing the lights and the cross bar, is equally distributed over both end-nodes ofthis beam.

For a better agreement with the actual test, the height at which in the 3-dimensional model thewing impacts the mast has been reduced with respect to the preliminary model, which wasthought to be more in accordance with the test case at the time. Further, the boundaryconditions of the ‘moving’ wing were modified in such a way, that the wing remains on thesame height (as in the actual test) during the whole analysis (by increasing the bendingstiffness). So, the wing can only deform in the longitudinal direction.

Measured failure criteria [6] were incorporated in the model for all beams. From severalpreliminary analyses with the 3-dimensional model it was observed that beams failed ratherearly in bending, which means that these beam were completely ‘removed’ in the analyses.The experimental results indicate, that after bending failure, the tubes still absorb energy dueto their remaining axial stiffness. To be able to take this into account in the model, theaffected beams were replaced with two beams, i.e., one which will fail in bending and onewhich will fail in axial tension or compression. These beams are located near the foundationof the mast, the connection with the wing and the connection with the masses of the landinglights at the top. In figures 3-4, these beams are indicated by the lines between the ‘opennodes’.

The model consists of 106 masses and 205 beams, including 2 non-linear beams (wing model)and 24 ‘double beams’ (to allow failure due to bending and due to tension/compression,separately).

ANALYSIS

The response of the approach light structure, due to an impact of the moving wing at an initialvelocity of 38.9 m/s (= 140 km/h), was analysed. This was simulated by the application of aninitial velocity of 38.9 m/s on the nodes, which represent the truck, the wing backing steelbeam, and the wing, respectively. The duration of the simulation was 40 milliseconds, with atime integration step size of 5.0 microseconds. This duration captures the phase in which mostof the damage to the wing section takes place.

Figure 5 shows the actual deformation mode as observed in test 1. The front vertical tubes cutthrough the wing nose, until they hit the wing spar. Subsequently, the mast moves with the

-7-NLR-TP-2000-618

wing, while the top and bottom of the mast remain in place. Finally, the aft vertical tubesfailed in tension at their foundation, but this happens at a later stage. In figure 6, thesimulation of the impact at several time steps up to 40 ms is presented, which shows a closeresemblance with the test results. Figure 7 shows, that the failure mode is not symmetrical,possibly due to the unsymmetrical layout of the diagonal tubes.

Figure 8 shows the velocity of several mass points. The mass points 103 and 106 represent thewing section, and start with an initial velocity of 40 m/s. This velocity reduces to 30 m/s,while in the end, the mass points will have to move with the truck, at an undisturbed 40 m/s.This point was not yet reached in the simulation. Mass point 23, representing (one half of) thetop mass, remains in place during the first 40 ms, like in the actual test (Fig. 5). Masspoint 18, just above the impact point, gradually picks up speed.

Figure 9 shows the forces in several beams. Beams 173 and 176 each represent one half of thetruck. Beams 174 and 177 each represent one half of the backing beam, and beams 175 and178 each represent one half of the wing section. The forces in the wing backing beam arethought to be most representative of the forces recorded in the test. These forces are added,and the sum is represented in figure 10. Also in figure 10, the test results are shown, andcompared to the numerical results. It is clear that the first two peak forces coincide well withthe experimental values, and also the maximum value is well predicted: 24 kN. Thesubsequent force history is less well predicted, but can be improved in the future, by using thecharacteristics of the wing impactor, as mentioned in chapter 3. In this respect, it was not yetrelevant to try to predict the impact energy, for comparison with the experimental value.

CONCLUSIONS

A computer code has been developed to analyse the impact problem of a light aircraft wingwith a frangible approach light structure. Hereto, the DRI/KRASH code was made availablein an extended form by DRI, according to the specifications of NLR. Subsequently, asimplified 2-dimensional model was analysed, to confirm the feasibility of this approach. Inthe present paper, an improved 3-dimensional model was evaluated, which contained thecharacteristics of the benchmark problem: full scale test results provided by the Finishcompany EXEL. The important parameters to be evaluated are the failure mode, the peakimpact force, and the impact energy.

The result of this evaluation was successful, in so far that the prediction of the failure modeand the peak impact force were correct. As the characteristics of the wing impactor were notavailable in time (the impactor was unexpectedly made available by Transport Canada), themodel was not yet accurate enough to predict the full force history. Hence, it was not possibleto generate the impact energy for comparison with the experimental value. This improvementwill be carried out in the near future.

The merits of the new simulation capability were clearly shown. One approach to simulatethis impact phenomenon would have been to use an explicit finite element code, such asPAM-CRASH [8]. However, a model for such a generic finite element analysis approachmight typically contain 50,000 degrees of freedom, which needs to be analysed for a largenumber of time steps, typically 10,000. This capability is technically out of reach of thedesigners of frangible approach masts, and very expensive to use for other interested parties.

-8-NLR-TP-2000-618

On the other hand, the hybrid capability provided by the DRI/KRASH code, as demonstratedhere, is available on a pc, taking only a few minutes to run. However, a KRASH-modelconsists of large elements of the size of full components, and requires test results for thesecomponents as input. As manufacturers usually perform component tests anyway during thedesign process, this effort should not be prohibitive.

In the near future, the DRI/KRASH model of the EXEL mast will be improved as indicatedabove, and the validation of the code will be undertaken, by applying it to a different type ofapproach light system, made of aluminium.

ADDITIONAL RESULTS

Since the first issue of the present report, additional numerical results were obtained. Theseresults were presented at the 3rd KRASH user’s seminar, January 8-10 2001, Arizona StateUniversity, Tempe, Arizona, USA.

The results consist of an improved simulation of the impact event, in terms of peak force andabsorbed energy. The results are presented in figures 11-12, and show good agreement for thefirst 70 % of the impact duration.

ACKNOWLEDGEMENT

The authors are pleased to acknowledge that the work herein was supported by the RLD(Netherlands Department of Civil Aviation), monitored by Mr. A.L. Franssen. They are alsograteful to EXEL Oy for making available the results of the full scale impact tests on theirmasts, and to Transport Canada to supply wing impactors for testing at NLR.

REFERENCES

[1] DRI/KRASH theory reference manual KRASH version 9601, Dynamic Response Inc.,January 1996.

[2] M. Luetzenburger, MLS, KRASH fore/post-buckling software user guide, MLS, 1998.[3] R.H.W.M. Frijns, G. Wittlin, A. de Boer, A preliminary model to simulate the impact of a

frangible approach light structure by an aircraft wing section, InternationalCrashworthiness Conference, IJCRASH ’98, 9-11 September 1998, Michigan, USA.

[4] J. Hanka, M. Vahteri, Impact tests of EXEL approach light masts – test report NESTE,October 1991.

[5] R.H.W.M. Frijns, J.F.M. Wiggenraad, Force data of soft impact tests on frangibleapproach light structures of EXEL performed in 1991, NLR CR 99491, 1999.

[6] R.H.W.M. Frijns, Structural tests to determine the mechanical properties of a frangibleapproach light mast made by Exel Oy, NLR CR 98341, 1998.

[7] R.H.W.M. Frijns, J.F.M. Wiggenraad, Static compression tests and computer models ofwing impactors used for impacts on frangible approach light towers, NLR CR 99495,1999.

[8] J.F.M. Wiggenraad, A.L.P.J. Michielsen, D. Santoro, F. Lepage, C. Kindervater,F. Beltran, Development of a crashworthy composite fuselage structure for a commuteraircraft, NLR TP 99532, 1999.

-9-NLR-TP-2000-618

a) Wing impactor mounted on truck b) Full scale impact at 140 km/h

c) Wing impactor after the test

Figure 1 Full scale impact test on frangible approach light structure

-10-NLR-TP-2000-618

Figure 2 Wing impactor and load cells

-11-NLR-TP-2000-618

a) Perspective b) Side view

c) Frontal view d) Top view

Figure 3 Different views of the impactor/mast model

-12-NLR-TP-2000-618

Figure 4 Details of model

-13-NLR-TP-2000-618

Figure 5 Photographic sequence of the impacted mast from EXELmast test 1

-14-NLR-TP-2000-618

Fig

ure

6 D

efor

mat

ion

of th

e m

ast,

com

pute

d w

ith

KR

ASH

-15-NLR-TP-2000-618

Fig

ure

6 D

efor

mat

ion

of th

e m

ast,

com

pute

r w

ith

KR

ASH

(con

tinu

ed)

-16-NLR-TP-2000-618

a) Perspective b) Side view

c) Frontal view d) Top view

Figure 7 Deformation of the mast from different views

Figure 8 Velocity of mass points

-17-NLR-TP-2000-618

Figure 9 Force versus time plots for the beams representing the truck, the wing backing beamand the wing section

-18-NLR-TP-2000-618

Figure 9 Force versus time plots for te beams representing the truck, the wing backing beamand the wing section (continued)

-19-NLR-TP-2000-618

Figure 10 Comparison of impact forces determinated by tests and analysis

-20-NLR-TP-2000-618

Figure 11 Force versus time

Figure 12 Impact energy versus time